Available online at www.sciencedirect.

com

ScienceDirect
Procedia CIRP 50 (2016) 390 – 395

26th CIRP Design Conference

A case study on the capability of rapid tooling thermoplastic laminating

moulds for manufacturing of CFRP components in autoclaves
Mario Lušića*, Kilian Schneidera, Rüdiger Hornfecka
a
Technische Hochschule Nürnberg, Institue for Chemistry, Materials and Product Development, Keßlerplatz 12, 90489 Nürnberg, Germany
* Corresponding author. Tel.: +49-911-5880-1906; fax: +49-911-5880-5900. E-mail address: mario.lusic@th-nuernberg.de

Abstract

Additive manufacturing (AM) is most used in prototype production and other processes upstream of series production. However,
in recent years additive manufacturing has also moved into series production environments substituting for established systems.
With AM, it is possible to improve production in terms of material consumption, manufacturing costs, and lightweight design.
Accordingly, it is possible to create complexly arched laminating moulds directly from the CAD-model instead of milling them
from solid material as is frequently done for the production of carbon fibre reinforced plastic (CFRP) components in autoclaves.
This work analyses the potential of CFRP-laminating moulds as rapid tooling moulds generated by fused deposition modeling. A
rounded cuboid will be considered with different reinforcement patterns as well as various wall thicknesses. Normal autoclave
conditions will be simulated with pressure variation and high temperature stress varying over time. In conclusion, the results
prove the capability of rapid tooling thermoplastic laminating moulds for manufacturing CFRP components in autoclaves.
© 2016
© 2016TheThe Authors.
Authors. Published
Published by Elsevier
by Elsevier B.V. This isB.V.
an open access article under the CC BY-NC-ND license
Selection and peer-review under responsibility of Professor Lihui Wang.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of the 26th CIRP Design Conference
Keywords: rapid tooling; moulding; autoclave; fused deposition modeling.

1. Introduction stiffness, which can lead to a part’s inability to withstand

Additive manufacturing [1,2] (AM) is most used in
prototype production and other processes upstream of series
production. Recently, AM has also moved into series
production environments where it is replacing established
systems. AM makes it possible to improve conventional
production in terms of material consumption, manufacturing
costs, and lightweight design. Accordingly, it is possible to
create complexly arched laminating moulds directly from
CAD-models instead of milling them from solid material as is
frequently done for the production of carbon fibre reinforced
plastic (CFRP) components in autoclaves. However,
mechanical design is still driven by producing components
with traditional manufacturing processes [3].
This leads to additive manufactured moulds that are
conservatively designed, which means they are composed
fully of solid material. However, AM permits hollowing out
the core of the part for reasons of lightweight design and thus
minimises manufacturing time as well as material
consumption. A major limitation to this process is the loss in
loads in subsequent manufacturing processes. [4]
2. Challenges in AM-generated CFRP moulds
The potential and limitations of AM-generated CFRP
laminating moulds as rapid tooling moulds remain to be more
clearly defined. In our previous case study [4], the potential
towards zero waste in the additive manufacturing of such
moulds was proven. However, that investigation focused on
meeting manufacturing loads by vacuum bag moulding [5]
and therefore only a manufacturing load of 1 bar ambient
pressure was investigated. In conclusion, those results are not
transferable to other common CFRP-manufacturing processes
as they do not reflect load variation, e.g., in temperature and
pressure. The present investigation will add to our earlier
results, taking into account the manufacturing loads caused by
an autoclave [7] during CFRP manufacturing. The main
criteria investigated in this study, in contrast to the previous
study, are:

2212-8271 © 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of the 26th CIRP Design Conference
doi:10.1016/j.procir.2016.04.151
 Mario Lušić et al. / Procedia CIRP 50 (2016) 390 – 395 391

x temperature over the ambient, expansion = 47 µm/(m°·C)). Since the mould is located in
x surrounding pressure loads higher than 1 bar, practice on a base (e.g., on a pin-type mould [8] or on a
(both temperature and pressure loads arising from the workbench), the ambient pressure p is exerted on its visible
conditions inside an autoclave) surface, see Fig. 2, but not on its base surface. Thus, the nodes
x and a common but heat-resistant AM-material, of the finite elements within the base surface cannot move
vertically, but can move horizontally.
besides the differences in the geometry of the test body [4,6],
see Table 1.
p p
Table 1. Comparison of investigation criteria.

criteria Galantucci et al. [6] Lušić et al. [4] this work

surrounding surrounding
load type swaging
pressure pressure
several loads
value of until reaching one atmospheric
given by an
load failure load load = 1bar Fig. 2. Ambient pressure on rapid tooling mould (test body with nominal
autoclave
dimensions in mm).
Polycarbonate by ABS-M30 by ULTEMTM 1010
material
Stratasys Stratasys by Stratasys In the same manner as described in [4], stair stepping or
anisotropic material properties common in AM are not
AM fused deposition fused deposition fused deposition
technology modeling modeling modeling considered here. The target values for this investigation are:

geometry cylinder twisted block symmetric block x maximum total deformation1,

x manufacturing time, and
solid material vs solid material vs
internal solid vs several hollowed vs cross hollowed vs cross x material consumption.
geometry narrow-waisted vs honeycomb vs honeycomb
structure structure 4. Experimental procedure
non-steady-state,
temperature not specified not specified This section addresses the experimental procedure using
over ambient
finite element analysis. The following test bodies were
stressed with temperature and pressure according to Fig. 1:
3. Model concept, constraints, and target values
x Firstly (section 4.1), a solid design was used, which
The model is based on an application scenario using an reflects reference values for deformation, manufacturing
autoclave [7] for manufacturing carbon fibre reinforced time, and material consumption.
plastics. So, the following boundary conditions were x Secondly (section 4.2), the test body becomes a shell by
determined for this parameter study (Fig. 1): hollowing out its inner core.
x Thirdly (section 4.3), an inner structure is simulated in two
x Ambient temperature: 20 °C to 120 °C. The heat transfer ways: once as a cross structure and once as a honeycomb
behaviour between the ambient air and the test body is not structure.
considered since we focused on deformation behaviour
during heat conduction within the rapid tooling mould. The test bodies described in sections 4.2 and 4.3 serve to
Thus, the temperature is assumed to be directly acting on reduce manufacturing time and material consumption while
the component’s surface. maintaining the maximum deformation values from the solid
x Ambient pressure: 1 to p bar, whereby p varies by design.
increments of 0.5 bar up to p = 3bar = pmax.
x Time: ambient temperature and pressure are not stationary 4.1. Determining reference values for the solid material
over time to reflect the process within an autoclave.
The maximum total deformation values for the different
140
120 pressure levels was about 0.28 mm directed outwards and was
temperature [°C]

pressure [bar]

100 pmax
80 reached in all cases at 120°C. In all other cases (shell design,
60
40
pressure cross-structure design, and honeycomb design, see sections
temperature
20 4.2, 4.3.1, and 4.3.2), the deformation was directed towards
0 1
0 25 50 75 85 100 125 135 150 the body centre2. With an increase in pressure the deformation
time [min]
increases, but not to a large extent. The maximum equivalent
Fig. 1. Profile of temperature and pressure used in this parameter study.
1
the equivalent stress is continually checked to confirm that it remains below
The material used is ULTEMTM 1010 (yield strength = 81 the yield strength of UltemTM1010.
2
MPa, elastic modulus = 3.5 x 10³ MPa, coefficient of thermal Keeping in mind that the stress is caused by combining temperature and
(over)pressure resulting in a deformation decrease by increased pressure.
392 Mario Lušić et al. / Procedia CIRP 50 (2016) 390 – 395

stress of the test body is expressed as approximately 9 MPa at

3.0 bar. In conclusion, the test body is slightly deformed as
well as slightly stressed at each pressure. This indicates
oversizing regarding material volume used within the test
bodies and implies further potential for minimising the
material needed. Fig. 3 illustrates the deformation behaviour at
3 bar.

Fig. 5. Maximum total deformation values depending on wall thickness and

Fig. 3. Determining the maximum total deformation [mm] using the
algorithm in the ANSYS Workbench 16.1 software.
As shown in Fig. 4, the required manufacturing time was
38 hours & 13 min and material consumption was about (3051
cm³ + 42 cm³ =) 3093 cm³. These values reflect the reference
values that are to be optimised with the investigations in the
following sections.
pressure using a shell design; determined by using the algorithms in the
ANSYS Workbench 16.1 software.
Table 2 expands and summarises the values at a pressure of
3 bar, as it is the pressure with the most deformation. A wall
thickness of 12 mm minimises the maximum total
deformation value of the shell to 0.28 mm, which is close to
the deformation value which results for solid design (see
section 4.1), but the shell is associated with about 57%3 lower
manufacturing time and 65%4 lower material consumption
than the solid design.

Table 2. Influence of wall thickness on all target (maximum) values at 3 bar

pressure.

wall thick- total deformation equivalent stress manufacturing time material consumption
ness in mm in mm in MPa in hr.min in cm³
3 25.54 130.87 7.18 ≈ 336

4 9.09 70.69 8.37 ≈ 424

5 4.58 46.63 9.44 ≈ 512

6 1.73 28.13 10.52 ≈ 596

7 1.33 20.46 11.45 ≈ 678

8 0.87 17.06 12.38 ≈ 761

9 0.63 12.06 13.32 ≈ 840

10 0.53 8.62 14.29 ≈ 924

11 0.36 6.55 15.22 ≈ 1001

Fig. 4. Determining the manufacturing costs using the algorithms in the
Insight 10.6 software by Stratasys.
4.2. Investigation of the influence of wall thickness on the
shell design
The idea of the following investigation is to create a cavity
inside the test body, thus saving material and reducing
manufacturing costs. To this end, the model is built as a shell
and the wall thickness varies by 1 mm increments between 3
mm and 10 mm. As shown in Fig. 5, the wall thickness plays a
more important role than the value of the pressure in
minimising the maximum total deformation, which decreases
disproportionately with increasing the wall thickness.
Increasing the pressure increases the maximum total
deformation, but to a greater extent for smaller than larger
wall thicknesses.
12 0.28 5.51 16.11 ≈ 1075
4.3. Investigation of designs with internal structures
The wall thickness and the design of the core structure
were defined as in our previous study [4], using a cross versus
a honeycomb structure (Fig. 6) standing perpendicularly on
the upper test body’s surface with following structural
variants [4]:
x thickness x: 1 to 5 mm varied by increments of 1 mm,
x distance z: 5 to 20 mm varied by 5 mm increments.
Compared to the different wall thickness in the shell’s
design in section 4.2, the value of the wall thickness is scaled
down and set to 2 mm for all core structure variations. This is
due to the fact that all internal structures lead to an increase in

3
((38 hours 13 min - 16 hours 11 min) / 38 hours 13 min) x 100% ≈ 57 %
4
((3093cm³ - 1075 cm³) / 3093 cm³ ) x 100 % ≈ 65 %
 Mario Lušić et al. / Procedia CIRP 50 (2016) 390 – 395 393

strength and therefore the wall thickness of the shell can be As shown in Fig. 9, the variable z has a more important
reduced to save material. influence on minimising manufacturing time than the variable
x. With an increase in z the manufacturing time decreases.
z x Whereas, Fig. 10 shows that variable x has a greater impact
on the material consumption compared to its influence on the
manufacturing time. Nonetheless, z has the more important
z

influence on reducing material consumption compared to x.

z

x
a) b)

Fig. 6. Core structures selected for the present investigation: a) cross

structure, b) honeycomb structure. [4]

4.3.1. Investigation of the cross structure

As shown in Fig. 7 and Fig. 8, with increasing value x, the

deformation decreases. The same occurs with decreasing Fig. 9. Manufacturing time for cross structure depending on variables x and z;
variable z. Deformation increases along with increasing determined by using the algorithms in Stratasys’ Insight 10.6.
pressure. However, this occurs to a lesser extent with
increased x and decreased z.
Only for the value x = 1 mm was the maximum
deformation value not reached for each pressure as given by
section 4.1, which is about 0.28 mm. In this case, the variable
z must not exceed 10 mm. But, already with the first
increment to x = 2 mm, the maximum deformation value
dropped under 0.2 mm for each pressure and, thus, clearly
under the reference value of 0.28 mm.
For value z, the reference deformation of less than 0.28
mm at maximum is reached for z = 5 and z = 10 at each
pressure and for each variable x.
Fig. 10. Material consumption for cross structure depending on variables x
and z; determined by using the algorithms in Stratasys’ Insight 10.6.

4.3.2. Investigation of the honeycomb structure

As shown in Fig. 11, the reference deformation of less than

or equal to 0.28 mm at maximum is already reached for x = 1
mm, and thus for all other x, and for each z.

Fig. 7. Maximum total deformation values (cross structure, x = 1 mm);

determined by using the algorithms in ANSYS’ Workbench 16.1.

Fig. 11. Maximum total deformation values (honeycomb structure, x = 1

mm); determined by using the algorithms in ANSYS’ Workbench 16.1.

It should be noted that with an increase in x there is no

Fig. 8. Maximum total deformation values (cross structure, x = 2 mm);
determined by using the algorithms in the ANSYS’ Workbench 16.1.
steady decrease in deformation for bodies with the
honeycomb structure, see Fig. 12 and Fig. 13. As illustrated
by Fig. 14, the deformation in the wall, where the maximum
deformation occurs, depends on the bending lengths, which in
turn depend on the value x. Thus, an increase in x can
394 Mario Lušić et al. / Procedia CIRP 50 (2016) 390 – 395

increase the bending length resulting in a deformation

increase.

Fig. 16. Material consumption for honeycomb structure depending on

variables x and z; determined by using the algorithms in Stratasys’ Insight
Fig. 12. Maximum total deformation values (honeycomb structure, x = 4 10.6.
mm); determined by using the algorithms in ANSYS’ Workbench 16.1.
5. Results and conclusions

As summarised in Table 3, this paper shows potentials for

minimising material use and manufacturing time needed while
maintaining stiffness in order to remain within deformation
limits while using rapid tooling thermoplastic laminating
moulds for manufacturing CFRP components in autoclaves.
All three test body designs required significantly less
manufacturing time and material consumption (≈ 65% / ≈ 74%
/ ≈ 84%) compared to the solid material mould design. In
Fig. 13. Maximum total deformation values (honeycomb structure, x = 5
conclusion, waste within additive manufacturing is reduced.
mm); determined by using the algorithms in ANSYS Workbench 16.1.
Table 3. Summarising manufacturing costs by test body design.

x = 3mm, z = 20mm x= 5mm, z = 20mm

manufacturing material material
test body
time consumption saving
solid material 38 hours 13 min ≈ 3093 cm³ ---
shell 16 hours 11 min ≈ 1075 cm³ ≈ 65%
(12 mm wall thickness)
cross structure 26 hours 24 min ≈ 815 cm³ ≈ 74%
(x = 2mm, z = 20 mm)

honeycomb structure 17 hours 36 min ≈ 493 cm³ ≈ 84%

Fig. 14. Different resulting lever arms/bending lengths (a < b) in the wall
depending on value x.
As shown in both Fig. 15 and Fig. 16, the same
dependency on x and z as described in section 4.3.1 for the
cross structure appears in the case of the honeycomb
structure.
Fig. 15. Manufacturing time of honeycomb structure depending on variables
x and z; determined by using the algorithms in Stratasys’ Insight 10.6.
(x = 1mm, z = 20 mm)
For the cross structure, with x = 2 mm, the variable z can
be increased to 20 mm for the purpose of minimising
manufacturing time and material consumption as much as
possible, but allowing a maximum deformation value of
clearly less than the reference value given in section 4.1. For
the honeycomb structure, this occurs even at x = 1 mm.
Comparing the optimal manufacturing costs for the cross
and honeycomb structures, the honeycomb structure shows
clearly more potential than the cross structure.
Comparing both cross and honeycomb structures with the
shell design, with a 12 mm wall thickness in order to conform
to the maximum deformation value of 0.28 mm given in
section 4.1, there is a reduction of material consumption by
about 24%5 for the cross structure and 54%6 for the
honeycomb structure. Although the manufacturing time
increases when the test body contains a core structure, the
increase is only slight for designs containing a honeycomb
structure.

5
((1075 cm³ - 815 cm³) / 1075 cm³) x 100% ≈ 24 %
6
((1075 cm³ - 493 cm³) / 1075 cm³) x 100% ≈ 54 %
 Mario Lušić et al. / Procedia CIRP 50 (2016) 390 – 395
6. Future work
Each AM process uses primarily materials designed
specifically for its technology. The properties of these
materials do not fully match the properties of materials used
within more common manufacturing processes, e.g., in plastic
injection moulding or casting. In particular, the layered
structure of additive manufacturing, which was not
investigated within this work, contributes to these differences.
Thus, it must be examined, how the direction-dependent
properties affect, e.g., the mechanical behaviour or the
coefficient of expansion of UltemTM 1010.
Furthermore, the coefficient of expansion was kept
constant in this investigation, so temperature dependences as
well as non-linearity were not considered.
As described in our previous study [4], bio-inspired
solutions for core structures besides the cross and honeycomb
structures provide further opportunity for developing
lightweight designs. Thus, implementing algorithms that build
up biological structures within additive manufactured
laminating moulds are desirable.
Acknowledgements
This report is a part of a research project funded by the
German Federal Ministry of Education and Research (funding
395
number: 03FH041PA3). The responsibility for the content of
this publication lies with the authors.
References
[1] Gibson I, Rosen D, Stucker B. Additive Manufacturing Technologies. 3D
Printing, Rapid Prototyping, and Direct Digital Manufacturing. New
York: Springer; 2015.
[2] Gebhardt A. Understanding Additive Manufacturing. Rapid Prototyping,
Rapid Tooling, Rapid Manufacturing. Munich: Carl Hanser; 2011.
[3] Gao W, Zhang Y, Ramanujan D, et al. The status, challenges, and future
of additive manufacturing in engineering. Computer-Aided Design (2015)
69, p. 65-89.
[4] Lušić M, Barabanov A, Morina D, Feuerstein F, Hornfeck R. Towards
zero waste in additive manufacturing: a case study investigating one
pressurised rapid tooling mould to ensure resource efficiency. Procedia
CIRP (2015) 37. p. 54-58.
[5] Ball P. Manufacturing Processes. In: Hollaway L, editor. Handbook of
polymer composites for engineers. Abington: Woodhead Publishin; 1994.
p. 73-98.
[6] Galantucci LM, Lavecchia F, Percoco G. Study of compression properties
of topologically optimized FDM made structured parts. CIRP Annals -
Manufacturing Technology 57 (2008), p. 243-246.
[7] Ball P. Manufacturing Processes. In: Hollaway L, editor. Handbook of
polymer composites for engineers. Abington: Woodhead Publishin; 1994.
p. 73-98.
[8] Lušić M, Wimmer M, Maurer C, Hornfeck R. Engineering framework for
enabling mass customisation of curvilinear panels with large surfaces by
using pin-type tooling. Procedia CIRP (2015) 37. p. 265-270.